Human neuroblastoma cell lines sometimes comprise two distinctly different morphological types of cells. One cell type (termed N) is small, has neuritic processes and adheres poorly to the substrate. The other is large, flat, and highly substrate-adherent (termed S). Karyotype analysis indicates that N and S cells from any one cell line arise from a common precursor. Moreover, as shown for the SK-N-SH cell line, N and S cells can undergo bidirectional morphological and biochemical interconversion. Whereas N cells express enzyme activities of noradrenergic neurons, S cells (of the SK-N-SH line) do not contain these enzymes, but rather, express activity for the melanocyte enzyme tyrosinase. New studies show that N and S cells also can be distinguished by expression of cell surface antigens revealed by serological analysis with a panel of monoclonal antibodies. Further, N and S cells show differential expressions of collagen, neurofilament protein, vimentin, and fibronectin. These results thus indicate that human neuroblastoma cells in culture undergo spontaneous bidirectional interconversion between cellular phenotypes (neuronal, melanocytic, glial) within the neural crest repertoire. This transdifferentiation phenomenon will be explored by continued isolation of N and S clones from a large group of neuroblastoma lines and characterization of clones as to 1) karyotype; 2) marker enzymes of neural crest cell derivatives; 3) cell surface antigen expression (mixed hemadsorption assays with a monoclonal antibody panel and EGF receptor binding assays); and 4) cytoskeletal and extracellular matrix proteins [neurofilaments, vimentin, GFAP, collagen(s), fibronectin]. The mechanisms of interconversion will be studied by assessing changes in the transcription and/or translation of oncogenes N-myc and c-src, EGF receptor, collagen isotypes and neurofilaments. Nude mice will be used to determine tumorigenic and invasive potentials of the two cell types. Differences between N and S cells in response to cancer chemotherapeutic agents will be investigated. The ultimate objective is to assess and exploit the natural phenotypic heterogeneity exhibited by human neuroblastoma cells and to understand the coordinate program of phenotypic transdifferentiation in terms of both malignant transformation and cell differentiation.